[ejabberd] Re: Cluster Help
bill.brazell at cingular.com
Thu Apr 7 01:17:20 MSD 2005
Bill Brazell <bill.brazell <at> cingular.com> writes:
Nuking the databases and rebuilding got it working. Thanks to all that helped.
Now I need to figure out a strategy regarding the tables and what settings to
give them. Anyone have any insight on that?
Is the best way to make an eJabber cluster completely fault-tolerant by adding a
load balancer such as the ones from F5 or Alteon? Without a load balancer, if a
user's "home" server went down, they couldn't log in. But with the shared
database, reviving that server would be easy.
Am I missing something here?
More information about the ejabberd